CN114974253A - Natural language interpretation method and device based on character image and storage medium - Google Patents

Abstract

The invention provides a natural language interpretation method, a device and a storage medium based on a character portrait, wherein the natural language interpretation method based on the character portrait comprises the following steps: receiving first voice data; determining a first interaction scene to which the first voice data belongs; selecting a first interaction behavior interpretation model corresponding to the first interaction scenario; performing voice recognition on the first voice data to obtain a first recognition result; and correcting the first key words meeting preset conditions in the first recognition result by using the first interactive behavior interpretation model. By determining the first interactive behavior interpretation model corresponding to the first interactive scene and correcting the first keyword in the first recognition result of the first voice data by using the first interactive behavior interpretation model, the efficiency of voice recognition/language interpretation is improved, and meanwhile, the accuracy of voice recognition/language interpretation can also be improved.

Description

A natural language interpretation method, device and storage medium based on portraits

technical field

The present invention relates to the technical field of speech recognition, in particular to a natural language interpretation method, device and storage medium based on person portraits.

Background technique

With the continuous development of computer technology, the methods of human-computer interaction are becoming more and more diversified and intelligent. At present, more and more interactive platforms adopt the method of voice interaction. Voice interaction can improve user interaction efficiency and Enhancing the fun has become an important way of human-computer interaction. For example, the self-service voice customer service system first throws questions to the user through voice, and then the user answers by voice. Another example is some navigation systems, shopping systems, etc., which require users to issue voice commands to control the displayed content. In these scenarios, it is necessary to accurately recognize the user's voice in order to give correct feedback.

However, the accuracy of the existing speech recognition methods or language interpretation methods is not high, and they cannot be interpreted and recognized based on person portraits.

SUMMARY OF THE INVENTION

Based on the above problems, the present invention proposes a natural language interpretation method, device and storage medium based on person portraits. By determining the first interaction behavior interpretation model corresponding to the first interaction scene, and using the first interaction behavior interpretation model Modifying the first keyword in the first recognition result of the first speech data can improve the efficiency of speech recognition/language interpretation and at the same time improve the accuracy of speech recognition/language interpretation.

In view of this, one aspect of the present invention proposes a natural language interpretation method based on portraits, including:

receiving first voice data;

determining the first interaction scene to which the first voice data belongs;

selecting a first interaction behavior interpretation model corresponding to the first interaction scene;

performing speech recognition on the first speech data to obtain a first recognition result;

Using the first interaction behavior interpretation model to correct the first keywords that meet the preset conditions in the first recognition result;

Wherein, the first interaction behavior interpretation model includes an association relationship between interaction scene information and character portrait information.

Optionally, after the step of determining the first interaction scene to which the first voice data belongs, the method further includes:

According to the first interaction scenario, prompting the user to make a first action and/or prompting the user to speak the first text data and/or prompting the user to input the first selection data;

collecting the first motion data and/or the first text data and/or the first selection data;

The first key information is extracted from the first action data and/or the first text data and/or the first selection data.

Optionally, the step of selecting a first interaction behavior interpretation model corresponding to the first interaction scene includes:

Based on the correspondence between the pre-established key information and the interaction behavior interpretation model, selecting an interaction behavior interpretation model that matches the first key information;

It is determined that the interaction behavior interpretation model matching the first key information is the first interaction behavior interpretation model corresponding to the first interaction scene.

Optionally, the step of using the first interaction behavior interpretation model to revise the first keywords that meet the preset conditions in the first recognition result includes:

extracting a first keyword that satisfies a preset condition from the first identification result;

extracting a first character portrait from the first interactive behavior interpretation model;

The first keyword is modified according to the first character portrait.

Optionally, the step of determining the first interaction scene to which the first voice data belongs includes:

extracting first attribute information from the first voice data;

The first interaction scene is determined according to the first attribute information.

Optionally, the first attribute information includes: a collection tool, collection method, collection time, collection location, number of people, and semantic environment of the first voice data.

Optionally, after the step of using the first interaction behavior interpretation model to revise the first keyword that meets the preset condition in the first recognition result, the method further includes:

outputting the first correction result for the first keyword in the form of voice or text;

receiving the user's evaluation feedback on the correction result;

When the evaluation feedback is a positive value, increase the priority of the first interaction behavior interpretation model corresponding to the first interaction scene;

When the evaluation feedback is a negative value, the priority of the first interaction behavior interpretation model corresponding to the first interaction scene is lowered.

Optionally, before the step of receiving the first voice data, the method further includes:

Determine the relationship between the first user and the second user from the user group, and generate a first relationship label by using the respective unique identity labels of the first user and the second user;

acquiring first interaction behavior data between the first user and the second user;

According to the first interaction behavior data and the first relationship label, construct a first persona portrait of the first user, a second persona portrait of the second user, and the first user and the first persona Interactive behavior database between two users;

Repeat the foregoing operations until all users have established character portraits according to different roles, and an interactive behavior database has been established between different character portraits, and extracted key information from the interactive behavior database;

Inputting the data in the interactive behavior database into a trained neural network to obtain multiple interactive behavior interpretation models;

A corresponding relationship between the key information and the plurality of interaction behavior explanation models is established.

Another aspect of the present invention provides a natural language interpretation device based on person portrait, comprising: a speech receiving module, a processing module, a speech recognition module and a result correction module;

the voice receiving module for receiving the first voice data;

The processing module is configured to determine a first interaction scene to which the first voice data belongs, and select a first interaction behavior interpretation model corresponding to the first interaction scene;

the voice recognition module, configured to perform voice recognition on the first voice data to obtain a first recognition result;

the result correction module, configured to use the first interaction behavior interpretation model to correct the first keywords that meet the preset conditions in the first recognition result;

Wherein, the first interaction behavior interpretation model includes an association relationship between interaction scene information and character portrait information.

A third aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores at least one instruction, at least one piece of program, code set or instruction set, the at least one instruction, the at least one piece of program , the code set or the instruction set is loaded and executed by the processor to implement the natural language interpretation method based on the person portrait as described in any one of the foregoing.

By adopting the technical solution of the present invention, the natural language interpretation method based on person portrait includes: receiving first voice data; determining a first interaction scene to which the first voice data belongs; selecting a first interaction scene corresponding to the first interaction scene A behavior interpretation model; performing speech recognition on the first speech data to obtain a first recognition result; and using the first interactive behavior interpretation model to revise the first keywords that meet preset conditions in the first recognition result. By determining the first interaction behavior interpretation model corresponding to the first interaction scene, and using the first interaction behavior interpretation model to correct the first keyword in the first recognition result of the first speech data, it is possible to improve speech recognition/ While improving the efficiency of language interpretation, it can also improve the accuracy of speech recognition/language interpretation.

Description of drawings

1 is a flowchart of a natural language interpretation method based on a portrait provided by an embodiment of the present invention;

FIG. 2 is a flowchart after the step of determining the first interaction scene to which the first voice data belongs in another embodiment of the present invention;

3 is a specific execution flow chart of a step of selecting a first interaction behavior interpretation model corresponding to the first interaction scene in another embodiment of the present invention;

4 is a specific execution flow chart of a step of using the first interaction behavior interpretation model to correct the first keyword in the first recognition result that satisfies a preset condition in another embodiment;

5 is a flow chart after the step of correcting the first keyword that satisfies a preset condition in the first recognition result by using the first interaction behavior interpretation model in another embodiment;

6 is a flowchart of a method for constructing an interactive behavior interpretation model in another embodiment;

FIG. 7 is a schematic block diagram of a natural language interpretation apparatus based on a person portrait provided by an embodiment of the present invention.

Detailed ways

In order to understand the above objects, features and advantages of the present invention more clearly, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments may be combined with each other in the case of no conflict.

Many specific details are set forth in the following description to facilitate a full understanding of the present invention. However, the present invention can also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited by the specific implementation disclosed below. example limitations.

The terms "first", "second" and the like in the description and claims of the present application and the above drawings are used to distinguish different objects, rather than to describe a specific order. Furthermore, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units is not limited to the listed steps or units, but optionally also includes unlisted steps or units, or optionally also includes For other steps or units inherent to these processes, methods, products or devices.

Reference herein to an "embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the present application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor a separate or alternative embodiment that is mutually exclusive of other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

The following describes a method, device, and storage medium for natural language interpretation based on portraits according to some embodiments of the present invention with reference to FIGS. 1 to 7 .

As shown in FIG. 1, an embodiment of the present invention provides a natural language interpretation method based on portraits, including:

receiving first voice data;

determining the first interaction scene to which the first voice data belongs;

selecting a first interaction behavior interpretation model corresponding to the first interaction scene;

performing speech recognition on the first speech data to obtain a first recognition result;

Using the first interaction behavior interpretation model to correct the first keywords that meet the preset conditions in the first recognition result;

Wherein, the first interaction behavior interpretation model includes an association relationship between interaction scene information and character portrait information.

It can be understood that a natural language interpretation method based on person portraits provided by the embodiments of the present invention can be applied to smart terminals, such as smart phones, computers, smart TVs, etc., and can also be used for intercom equipment, robots, access control systems, etc. .

In the embodiment of the present invention, the first voice data may be acquired through a voice acquisition unit (eg, a microphone), or the first voice data may be acquired from a server or an intelligent terminal through a communication network. During the process of collecting the first voice data, the relevant information of the voice occurrence scene is stored as the first attribute information of the first voice data at the same time.

It should be noted that, after the first voice data is received, the first interaction scene to which the first voice data belongs can be determined according to the first attribute information carried by the first voice data. The first attribute information of the voice data is the collection location, and the building (such as home, company, shopping mall, etc.) corresponding to the coordinates can be determined by the coordinates of the collection location; assuming that the collection location is a company, combined with other first attribute information, such as the collection time (such as the week of the week) 10:00 a.m.), the number of people (if there are 5 people; it can be determined by the voiceprint feature), etc., the interaction scene to which the first voice data belongs may be a "company meeting". According to actual application scenarios, the interaction scenarios may include but are not limited to: family chat, work discussion, shopping, friends gathering, etc.

Further, a first interaction behavior interpretation model corresponding to the first interaction scene is selected. Wherein, the first interaction behavior interpretation model includes the relationship between the interaction scene information and the character portrait information. In this way, the first recognition result can be interpreted or corrected in a targeted manner by determining the corresponding character portrait information.

In the embodiment of the present invention, to perform speech recognition on the first voice data, the first voice data may be segmented according to different voiceprints by a voice recognition module, or the first voice data may be divided according to a preset time length. The first voice data is segmented, and the first voice data can also be segmented according to a preset file size, and each segment of the segmented voice is queued according to the chronological sequence of speech occurrences, and a speech recognition algorithm is used. Each speech segment is converted into corresponding text information according to the queuing sequence; the text information is fused in chronological order, and adjusted according to the context to obtain a first recognition result.

For the first keywords (such as terms with local characteristics, industry common terms, professional terms, etc.) that meet preset conditions (such as occurrence frequency and/or error-prone frequency in the preset range) in the first recognition result, use The first interactive behavior interpretation model interprets/amends the first keyword to obtain a first interpretation/correction result, for example, using the character portrait information contained in the first keyword to interpret/amend the first keyword in combination with the character's characteristics.

Using the technical solution of this embodiment, the natural language interpretation method based on the person portrait includes: receiving first voice data; determining a first interaction scene to which the first voice data belongs; selecting a first interaction scene corresponding to the first interaction scene Interaction behavior interpretation model; perform speech recognition on the first voice data to obtain a first recognition result; use the first interaction behavior interpretation model to revise the first keywords that meet preset conditions in the first recognition result . By determining the first interaction behavior interpretation model corresponding to the first interaction scene, and using the first interaction behavior interpretation model to interpret/correct the first keyword in the first recognition result of the first voice data, improve the voice While the efficiency of recognition/language interpretation can be improved, the accuracy of speech recognition/language interpretation can also be improved.

As shown in FIG. 2, in some possible embodiments of the present invention, after the step of determining the first interaction scene to which the first voice data belongs, the method further includes:

According to the first interaction scenario, prompting the user to make a first action and/or prompting the user to speak the first text data and/or prompting the user to input the first selection data;

collecting the first motion data and/or the first text data and/or the first selection data;

The first key information is extracted from the first action data and/or the first text data and/or the first selection data.

It should be noted that, in order to further clarify the characteristics of the user's interaction behavior in the first interaction scenario, so as to select the optimal interaction behavior interpretation model, in the embodiment of the present invention, according to the first interaction scenario, some An interaction event to obtain the interaction behavior of the user in the first interaction scenario, for example, issuing an instruction to prompt the user to make a first action, and/or providing a piece of first text data and issuing an instruction to prompt the user to speak the first text data, and/or Provide some options and issue instructions prompting the user to enter first selection data, etc. By collecting the first action data and/or the first text data and/or the first selection data, and extracting the first key information therefrom. The first key information may be a specific action, a special tone, a special pronunciation, or a special preference, etc., which is not limited by the embodiments of the present invention.

It can be understood that, based on the first interaction scenario, the more interaction events are constructed and the more comprehensive the types of events covered, the more interaction behavior data is obtained, and the more accurate the interaction behavior interpretation model is selected later.

As shown in FIG. 3, in some possible embodiments of the present invention, the step of selecting a first interaction behavior interpretation model corresponding to the first interaction scene includes:

Based on the correspondence between the pre-established key information and the interaction behavior interpretation model, selecting an interaction behavior interpretation model that matches the first key information;

It is determined that the interaction behavior interpretation model matching the first key information is the first interaction behavior interpretation model corresponding to the first interaction scene.

It can be understood that, in the embodiment of the present invention, by analyzing the interactive behavior data in different historical interaction scenarios, and after processing by the neural network, the correspondence between the key information and the interactive behavior interpretation model is established. relation. As mentioned above in the explanation of the first key information, the key information may be a specific action, a special tone, a special pronunciation, or a special preference, and the like. Based on the correspondence between the key information and the interaction behavior interpretation model, the interaction behavior interpretation model matching the first key information is selected as the first interaction behavior interpretation model corresponding to the first interaction scene. Through this solution, the first interaction behavior interpretation model can be quickly and accurately selected, the execution efficiency is improved, and the user experience is improved.

As shown in FIG. 4 , in some possible implementations of the present invention, the step of using the first interaction behavior interpretation model to correct the first keyword that meets the preset condition in the first recognition result includes:

extracting a first keyword that satisfies a preset condition from the first identification result;

extracting a first character portrait from the first interactive behavior interpretation model;

The first keyword is modified according to the first character portrait.

It can be understood that, in this embodiment of the present invention, the first recognition result may be text information, and the first recognition result is extracted from the first recognition result that satisfies a preset condition (for example, the frequency of occurrence and/or the error-prone frequency is within a preset range. ) of the first keyword (such as terms with local characteristics, industry common terms, professional terms, etc.); then extract the first character portrait from the first interactive behavior interpretation model, and use the first character portrait contained in the Character tags (such as habitual residence, industry, accent characteristics, gender, character relationship, etc.), comprehensively analyze the first recognition result, and when there is an error, interpret/correct the first keyword to obtain the first Interpret/correct results. In this embodiment, the first recognition result is analyzed in a targeted manner by using the persona portraits, and the first keyword is corrected, which greatly improves the recognition readiness rate.

In some possible implementations of the present invention, the step of determining the first interaction scene to which the first voice data belongs includes:

extracting first attribute information from the first voice data;

The first interaction scene is determined according to the first attribute information.

It can be understood that, as mentioned above, in the process of collecting the first voice data, the relevant information of the voice occurrence scene is stored as the first attribute information of the first voice data, specifically, the voice data is stored. Packing with the first attribute information to form the first voice data, or modifying the data format of the voice data, adding a part to record the first attribute information to form the first voice data.

Wherein, the first attribute information includes: a collection tool of the first voice data (such as a mobile phone, a drone, an office robot, a smart camera, etc.), a collection method (such as direct collection through a device, collection through a network connection to other devices, etc.) etc.), collection time (such as 6 am, 9 am, 3 pm, 8 pm, etc.), collection location (such as company, home, shopping mall, hospital, school, etc.), number of characters and semantic environment (mainly including expression , comprehension of foreword and context).

In the embodiment of the present invention, by recording the relevant information of the speech occurrence scene, an additional reference dimension is provided for the subsequent speech recognition, and the speech recognition efficiency and accuracy are improved.

As shown in FIG. 5 , in some possible implementations of the present invention, after the step of using the first interaction behavior interpretation model to correct the first keywords that meet the preset conditions in the first recognition result ,Also includes:

outputting the first correction result for the first keyword in the form of voice or text;

receiving the user's evaluation feedback on the correction result;

When the evaluation feedback is a positive value, increase the priority of the first interaction behavior interpretation model corresponding to the first interaction scene;

When the evaluation feedback is a negative value, the priority of the first interaction behavior interpretation model corresponding to the first interaction scene is lowered.

It can be understood that, in order to improve the recognition efficiency, the embodiment of the present invention sets up a feedback mechanism, by outputting the first interpretation/correction result of the first keyword in the form of voice or text, and providing the user with the recognition result. An interactive interface for evaluation feedback; receiving the user's evaluation feedback on the correction result, and correspondingly adjusting the first interaction behavior interpretation model corresponding to the first interaction scene according to the positive and negative values represented by the evaluation feedback. priority.

As shown in FIG. 6, in some possible embodiments of the present invention, before the step of receiving the first voice data, the method further includes:

S1, determine the relationship between the first user and the second user from the user group, and generate a first relationship label by using the respective unique identity labels of the first user and the second user;

It can be understood that each user has a unique identity tag, the relationship between users can be role relationship such as parent, child, husband and wife, friend, colleague or other basic relationship, and the user's role tag can be based on its unique identity tag. The construction is performed through preset rules, such as adding a role field after the unique identity tag, and the construction of the first relationship tag may be a fusion of the role tags of the first user and the second user. In this step, any two different users are randomly selected as the first user and the second user.

S2, acquiring first interaction behavior data between the first user and the second user;

In this step, the interactive behavior includes chatting, discussion, teaching, commands, etc., from the interactive behavior data between multiple users/characters (such as voice, action, text, geographic location, character distance, number of simultaneous participants, background noise) etc.), extract the first interaction behavior data between the first user and the second user.

S3. According to the first interaction behavior data and the first relationship label, construct a first persona portrait of the first user, a second persona portrait of the second user, and the first user and all Describe the interactive behavior database between the second users;

In this step, character portraits are established by using words, emotions, age, gender, education stage, accent, hobbies, etc., based on character portraits and necessary technical means (keyword identification, emotion identification and attitude analysis, etc.) Establish an interactive behavior database between the characters, and the interactive behavior database contains the relationship label of each interactive behavior (as constructed from the aforementioned first relationship label);

S4, repeating the foregoing operations until all users have established character portraits according to different roles, and an interactive behavior database has been established between different character portraits, and extracted key information from the interactive behavior database;

In this step, the key information, such as the aforementioned first key information, may be a specific action, a special tone, a special pronunciation, or a special preference.

S5, inputting the data in the interactive behavior database into the trained neural network to obtain multiple interactive behavior interpretation models;

S6. Establish a corresponding relationship between the key information and the multiple interaction behavior interpretation models.

This embodiment is based on the role relationship between users and the interactive behavior data as the analysis basis, and uses the pre-trained neural network to generate and construct the interactive behavior interpretation model. The execution method is simple, and the generated interactive behavior interpretation model can also provide accurate information. Interpret the results.

In some possible implementations of the present invention, a verification mechanism is established, and by setting a plurality of corresponding interaction behavior interpretation models for the same interaction scene, the first keyword in the first recognition result of the first speech data is passed through multiple After the interactive behavior explanation model is explained/corrected, the user selects the most correct interpretation/correction result. Specifically, after using the first mutual behavior interpretation model to interpret/correct the first keywords that meet the preset conditions in the first recognition result, and obtain the first interpretation/correction result, the method further includes:

selecting a second interaction behavior interpretation model corresponding to the first interaction scene;

Using the second interactive behavior interpretation model to interpret/correct the first keywords in the first recognition result that meet the preset conditions, to obtain a second interpretation/correction result;

Display the first interpretation/correction result and the second interpretation/correction result to the user, and the user selects the best one, and adjusts the first interaction behavior interpretation model and the second interaction according to the user's selection The preference for behavioral explanation models.

As shown in FIG. 7 , another embodiment of the present invention provides a natural language interpretation device 700 based on person portraits, including: a speech receiving module 701, a processing module 702, a speech recognition module 703, and a result correction module 704;

The voice receiving module 701 is configured to receive first voice data;

The processing module 702 is configured to determine a first interaction scene to which the first voice data belongs, and select a first interaction behavior interpretation model corresponding to the first interaction scene;

The speech recognition module 703 is configured to perform speech recognition on the first speech data to obtain a first recognition result;

The result correction module 704 is configured to use the first interaction behavior interpretation model to correct the first keywords that meet the preset conditions in the first recognition result;

Wherein, the first interaction behavior interpretation model includes an association relationship between interaction scene information and character portrait information.

For the operation method of the apparatus provided in this embodiment, please refer to the foregoing method embodiments, and details are not described herein again.

FIG. 7 is a schematic diagram of the hardware composition of the apparatus in this embodiment. It will be appreciated that Figure 7 only shows a simplified design of the device. In practical applications, the device may also include other necessary elements, including but not limited to any number of input/output systems, processors, controllers, memories, etc., and all the elements that can implement the natural language interpretation method of the embodiments of the present application The devices are all within the scope of protection of the present application.

Another embodiment of the present invention provides a computer-readable storage medium, the computer-readable storage medium stores at least one instruction, at least one segment of a program, a code set or an instruction set, the at least one instruction, the at least one segment of A program, the set of codes or the set of instructions is loaded and executed by a processor to implement any of the methods described above.

It should be noted that, for the sake of simple description, the foregoing method embodiments are all expressed as a series of action combinations, but those skilled in the art should know that the present application is not limited by the described action sequence. Because in accordance with the present application, certain steps may be performed in other orders or concurrently. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

In the above-mentioned embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

In the several embodiments provided in this application, it should be understood that the disclosed apparatus may be implemented in other manners. For example, the device embodiments described above are only illustrative. For example, the division of the above-mentioned units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated. to another system, or some features can be ignored, or not implemented. On the other hand, the shown or discussed mutual coupling or direct coupling or communication connection may be through some interfaces, indirect coupling or communication connection of devices or units, and may be in electrical or other forms.

The units described above as separate components may or may not be physically separated, and components shown as units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution in this embodiment.

In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically alone, or two or more units may be integrated into one unit. The above-mentioned integrated units may be implemented in the form of hardware, or may be implemented in the form of software functional units.

The above-mentioned integrated units, if implemented in the form of software functional units and sold or used as independent products, may be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application can be embodied in the form of a software product in essence, or the part that contributes to the prior art, or all or part of the technical solution, and the computer software product is stored in a memory, Several instructions are included to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the above-mentioned methods in the various embodiments of the present application. The aforementioned memory includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

Those skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing relevant hardware through a program, and the program can be stored in a computer-readable memory, and the memory can include: a flash disk , Read-only memory (English: Read-Only Memory, referred to as: ROM), random access device (English: Random Access Memory, referred to as: RAM), magnetic disk or optical disk, etc.

The embodiments of the present application have been introduced in detail above, and the principles and implementations of the present application are described in this paper by using specific examples. The descriptions of the above embodiments are only used to help understand the methods and core ideas of the present application; at the same time, for Persons of ordinary skill in the art, based on the idea of the present application, will have changes in the specific implementation manner and application scope. In summary, the contents of this specification should not be construed as limitations on the present application.

Although the present invention is disclosed above, the present invention is not limited thereto. Any person skilled in the art, without departing from the spirit and scope of the present invention, can easily think of changes or substitutions, and can make various changes and modifications, including the combination of the above-mentioned different functions and implementation steps, including the implementation of software and hardware. The methods are all within the protection scope of the present invention.

Claims (10)

1. a natural language interpretation method based on portrait, is characterized in that, comprises: receiving first voice data; determining the first interaction scene to which the first voice data belongs; selecting a first interaction behavior interpretation model corresponding to the first interaction scene; performing speech recognition on the first speech data to obtain a first recognition result; Using the first interaction behavior interpretation model to correct the first keywords that meet the preset conditions in the first recognition result; Wherein, the first interaction behavior interpretation model includes an association relationship between interaction scene information and character portrait information. 2. The natural language interpretation method based on person portrait according to claim 1, characterized in that, after the step of determining the first interaction scene to which the first voice data belongs, further comprising: According to the first interaction scenario, prompting the user to make a first action and/or prompting the user to speak the first text data and/or prompting the user to input the first selection data; collecting the first motion data and/or the first text data and/or the first selection data; The first key information is extracted from the first action data and/or the first text data and/or the first selection data. 3. The natural language interpretation method based on person portrait according to claim 2, wherein the step of selecting the first interaction behavior interpretation model corresponding to the first interaction scene comprises: Based on the correspondence between the pre-established key information and the interaction behavior interpretation model, selecting an interaction behavior interpretation model that matches the first key information; It is determined that the interaction behavior interpretation model matching the first key information is the first interaction behavior interpretation model corresponding to the first interaction scene. 4. The natural language interpretation method based on person portrait according to claim 3, wherein the first keyword that satisfies a preset condition in the first recognition result is analyzed by using the first interactive behavior interpretation model. The steps to make corrections include: extracting a first keyword that satisfies a preset condition from the first identification result; extracting a first character portrait from the first interactive behavior interpretation model; The first keyword is modified according to the first character portrait. 5. The natural language interpretation method based on person portrait according to claim 4, wherein the step of determining the first interaction scene to which the first voice data belongs comprises: extracting first attribute information from the first voice data; The first interaction scene is determined according to the first attribute information. 6. The natural language interpretation method based on person portrait according to claim 5, wherein the first attribute information comprises: a collection tool, collection method, collection time, collection location, character of the first voice data Quantitative and semantic context. 7 . The natural language interpretation method based on person portrait according to claim 6 , wherein, in the first key that satisfies a preset condition in the first recognition result using the first interactive behavior interpretation model, After the steps of word correction, it also includes: outputting the first correction result for the first keyword in the form of voice or text; receiving the user's evaluation feedback on the correction result; When the evaluation feedback is a positive value, increase the priority of the first interaction behavior interpretation model corresponding to the first interaction scene; When the evaluation feedback is a negative value, the priority of the first interaction behavior interpretation model corresponding to the first interaction scene is lowered. 8. The natural language interpretation method based on person portrait according to claim 7, is characterized in that, before the described step of receiving the first speech data, also comprises: Determine the relationship between the first user and the second user from the user group, and generate a first relationship label by using the respective unique identity labels of the first user and the second user; acquiring first interaction behavior data between the first user and the second user; According to the first interaction behavior data and the first relationship label, construct a first persona portrait of the first user, a second persona portrait of the second user, and the first user and the first persona Interactive behavior database between two users; Repeat the foregoing operations until all users have established character portraits according to different roles, and an interactive behavior database has been established between different character portraits, and extracted key information from the interactive behavior database; Inputting the data in the interactive behavior database into a trained neural network to obtain multiple interactive behavior interpretation models; A corresponding relationship between the key information and the plurality of interaction behavior explanation models is established. 9. A natural language interpretation device based on portraits, comprising: a voice receiving module, a processing module, a voice recognition module and a result correction module; the voice receiving module for receiving the first voice data; The processing module is configured to determine a first interaction scene to which the first voice data belongs, and select a first interaction behavior interpretation model corresponding to the first interaction scene; the voice recognition module, configured to perform voice recognition on the first voice data to obtain a first recognition result; the result correction module, configured to use the first interaction behavior interpretation model to correct the first keywords that meet the preset conditions in the first recognition result; Wherein, the first interaction behavior interpretation model includes an association relationship between interaction scene information and character portrait information. 10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores at least one instruction, at least one piece of program, code set or instruction set, and the at least one instruction, the at least one piece of program, the code set or the instruction set is loaded and executed by the processor to Realize the natural language interpretation method based on person portrait as described in any one of claims 1 to 8.

Images